Revision of Eragrostis
Total Page:16
File Type:pdf, Size:1020Kb
BLUMEA 47 (2002) 157-204 Revision of Eragrostis (Gramineae, Chloridoideae) in Malesia J.F. Veldkamp Nationaal Herbarium Nederland,Universiteit Leiden branch P.O. Box 9514, 2300 RA Leiden, The Netherlands e-mail: [email protected] Summary In Malesia there are 25 taxa of Eragrostis (Gramineae,Chloridoideae,Eragrostideae), incl. Ectro- siopsis (Ohwi) Jansen. Eight, perhaps eleven species are introduced. Eragrostis balgooyi Veldk. is here described as new. Eragrostis amabilis (L.) Nees is the correct name for E. tenella (L.) Roem. & Schult. The combination E. elongata (Willd.) Jacq. has been applied to at least eight taxa ranging from Sri Lanka and the E. diandra is used here. to Australia, younger synonym, (R.Br.) Steud., Key words '. Ectrosiopsis, Eragrostis, Gramineae, Malesia. Introduction Wolf Chloridoideae, is of Eragrostis (Gramineae, Eragrostideae) a large genus grasses for its with c. 350 species in the (sub)tropics, notorious troublesome infrageneric and delimitation.The characters used, of modeof specific e.g. presence glands, fragmenta- tion of the spikelets, number and size of the anthers, shape of the caryopses, etc., are often difficult to observe and assess. Some closely related species, e.g. those of the E. atrovirens/E. ‘elongata ’-complex; are common weeds, but in the herbarium can only be distinguished after careful analysis of the spikelets under the microscope. Recently, some importantpapers havebeen published on the infrageneric delimita- tion (Amarasinghe & Watson, 1990;Van den Borre & Watson, 1994;Lazarides, 1997; Gomez Sanchez & Koch, 1998) to which is here referred for further information. There are also some recent local revisions to which the interested readeris referred: Graterol et al. (1989: Venezuela), Kami (1993: Congo), Lazarides (1997: Australia), Cope (1998: Flora Zambesiaca area), Nicora (1998: Argentina), and Portal (2002: W Europe). NOTES ON THE INFRAGENERIC TAXONOMY OF THE GENUS Traditionally the genus is divided on the basis of the disarticulationof the spikelet for historical summarised (see Van den Borre & Watson (1994) a overview, in their table In Malesia the are encountered & Renvoize, 1). following groups (after Clayton 1986; Lazarides, 1997). Their sequence here has no taxonomic implication. I have re- frained from disentangling their synonymy and typification, as this was beyond the scope of the present study. 158 BLUMEA Vol. 47, No. 1, 2002 1) Lemmas and paleas with their joints disarticulating from above downward (sect. Cataclastos (Doll) Benth. or Psilantha (K. Koch) Tzvelev); 2) Lemmasand paleas with theirjoints disarticulating fromabove downward, lemmas more or less aristate (sect. Ectrosiopsis Ohwi); 3) Lemmas disarticulating from below upwards, rhachillaand sometimes paleas per- sisting. Spikelets very tardily disarticulating in the cereal E. tef (selected for this feature; collected in Malesia only once very long ago, but included here because new attempts at cultivation may be made) (sect. Lappula Stapf); 4) Lemmas disarticulating frombelow upwards, paleas long-persistent, the rhachilla ultimately breaking up fromabove downward (sect. Eragrostis or Pteroëssa Doll). To the first the amabilis- which small natural Old World belongs E. group, seems a where the bears the keels. In Malesia one with c. 20 species palea long setae on we have: E. amabilis (L.) Nees (incl. E. tenella (L.) Roem. & Schult., E. viscosa (Retz.) Steud., P. Beauv., and E. Hack. Trin.), E. aspera (Jacq.) E. riparia (Willd.) warburgii Here the lemma, palea, and caryopsis sometimes disarticulate fromabove as a whole lemma unit (E. aspera, E. warburgii), sometimes in fragments with the and caryopsis falling off and the palea persisting for a while (E. amabilis, E. riparia). The second has a single species, E. lasiocladaMerr., peculiar for having mucronate to awned distal lemmas for which reason Ohwi (1947) considered it to represent a separate section, Ectrosiopsis, elevated to generic rank by Jansen (1952, 1953), but included in Ectrosia R.Br, by Blake (1969) and Lazarides (1997: 166). Analyses by Van den Borre & Watson (1997) indicate that it is a distinct genus, but the DELTA program using INTKEY suggests that there would only be a difference in the relative length of the internodes of the spikelet. The presence of mucronate to awned distal lemmas be in SE but these also in African may striking Asia, occur some species, e.g. E. dinteri Stapf, E. rogersii C.E. Hubb., and E. variegata Welw. For the timebeing I have retained the species in Eragrostis. Ectrosia differsfrom Ectrosiopsis by the spikelets that dehisce above the glumes, but not between the heteromorphous lemmas, by the basal bisexual and upper sterile florets, the length of the basal cell of the microhairs, and the structure of the bundle sheaths and abaxial epidermis of the blades. The third is represented by e.g. E. atrovirens (Desf.) Steud., E. luzoniensis Steud., E. multicaulis Steud., E. pilosa (L.) P. Beauv., E. tef(Zuccagni) Trotter, and E. unio- loides (Retz.) Steud. The fourthin Malesia includes the E. ‘elongata '-complex (with E. brownii (Kunth) Nees, E. cumingii Steud., and E. diandra(R.Br.) Steud., etc.), E. cilianensis (Bellardi) Janch., E. curvula (Schrad.) Nees, E. gangetica (Roxb.) Steud., E. minor Host, E. montana Balansa, E. nigra Steud., E. pectinacea (Michx.) Steud., and E. tenuifolia (A. Rich.) Steud. Here would also belong DiandrochloaDe Winter(incl. Roshevitzia Tzvelev) which has been distinguished for some species with a membranous ligule and two anthers, but this group has not been generally accepted. The only Malesian representative is E. japonica (Thunb.) Trin. The membranous ligule is curious, but the presence of two anthers as a generic character seems of doubtful validity, as this character state occurs in widely diverse J. F. Veldkamp: Revision of Eragrostis in Malesia 159 taxa as even the present small sample of species shows, and two and three anthers within and may occur a single species (E. cumingii E. warburgii). Van den Borre & Watson's analysis (1994) showed that these sections are unnatural and they divided the genus into two subgenera especially based on leafanatomy and photosynthesis type. In cladograms ofthe Chloridoideaeby Van den Borre & Watson (1997) these two turn up in different clades (e.g. their figure 6 and table 3), so two in different be involved; genera, perhaps even subtribes, appear to according to the in the would be taxonomy used present paperEctrosiopsis appear to the generic name availablefor Eragrostis subgen. Caesiae at the generic level. At present it is premature to propose new combinations, as apparently no clear-cut macromorphological character supports the division of Eragrostis making it impractical in floristic revisions as the present one (see Van den Borre & Watson (1994: 407, table 7)). Lazarides well-awareofthis nevertheless 6 informal (1997) study recognised groups for Australia based on spikelet dehiscence, which was admittedly not supported by the types of photosynthetic carbon reduction (PCR) known or postulated from the bundle sheath anatomy (Kranz and subtypes) and microhairtype (see his table 1 and accompanying discussion). nine informal for the Zambesiaca Cope (1998) distinguished groups Flora area. SOME NOMENCLATURAL PROBLEMS Eragrostis sect. Ectrosiopsis was validly described by Ohwi (1947) including a Latin diagnosis. Jansen (1952, 1953) in a series of papers on Malesiangrasses elevated it to generic level. In the first part the reference to the basionym was given and 4 species in the second another added. the latter were included, part was Unfortunately, appeared ' before the first. Because the heading was Ectrosiopsis (Ohwi) Jansen' an indirect reference to the 2000: Art. and the basionym was provided (Greuter, 32.3,4) as paper appeared on December 1, 1952, it was only just in time for the generic name to be validly published: Art. 33.3 requiring a full and direct reference becomes into effect for after mentionedis therefore publications on or January 1,1953. The single species valid as well. Even without the reference the names would have been valid as then a descriptio generico-specifica would have been present (Art. 42.1). Blake's remark (1969) that the generic and specific name would be invalid is therefore incorrect. He reduced the genus to a single species in Ectrosia, but used the wrong combination, E. eragrostoides Domin (1915), citing the older E. lasioclada first in its he correct Merr. (1906) as the name synonymy. In 1973 made the combi- nation, saying that "the line with the necessary new combination was accidentally omitted... in 1969". Eragrostis elongata was described by Willdenow(1809) as Poa elongata based on a plant cultivated in the Berlin Botanical Garden from seed thought to have come from India. Vienna this cultivated In other plants under name were and depicted by Jacquin (1811) as Eragrostis elongata. Here three anthers were shown andlater authors have the whole of 3-anthered in SE of applied epithet to a range species Asia, some which are very widely distributedfrom Indiato Australia, e.g. E. brownii, E. bulbillifera Steud., E. cumingii, E. pubescens (R. Br.) Steud., E. spartinoides Steud., E. subsecunda (Lam.) E. Fourn., or E. zeylanica Nees, either as distinct species, or in various combi- 160 BLUMEA Vol. 47, No. 1, 2002 nations of them. Without corroborating remarks or descriptions it is only possible from voucher specimens to deduce what the author had in mind. The situation became even more confusing when Hubbardexamined the holotype in B and discovered that it belonged to a nearly strictly Australian2(! )-anthered species, E. diandra. Some possible isotypes in L were similarly annotated by Blake. How an Australian species during the Napoleonic wars of the early 19th century, when the Continent blocked from from Albion', could have turned was any import 'perfidious in Berlinis in Australianliterature up an intriguing question. Subsequently, some mainly E. elongata has been used for E. diandra (e.g. by Lazarides, 1997), but as other taxa Australia it is to which the name has been misapplied occur in as well, usually not clear to what taxon the author has applied the combination.